The present invention relates to a microactuator used to drive optical components and small-size magneto-optical/magnetic disk components, and a method of manufacturing the same.
A microactuator (electrostatic actuator) is generally proposed in which a movable element made of an insulating substance is moved by an electrostatic force generated between a plurality of stationary electrodes and the charges induced by the movable element when a voltage is applied to the plurality of stationary electrodes opposing the movable element at a small gap.
A microactuator mounted at the distal end of a suspension supported by an arm in a magnetic disk apparatus to drive a magnetic head formed integrally with a slider is proposed in L. S. Fan et al., "Magnetic Recording Head Positioning at Very High Track Densities Using a Microactuator-Based, Two-Stage Servo System", IEEE Transactions on Industrial Electronics, Vol. 42, No. 3, pp. 222-233, June 1995 (reference 1).
FIG. 12 shows a microactuator described in reference 1.
In FIG. 12, the conventional microactuator is constituted by a pair of T-shaped stationary elements 83 and 84 which are formed on a silicon substrate (to be described later) and have the distal ends of leg portions opposing each other, and an H-shaped movable element 82 formed between the stationary elements 83 and 84. The movable element 82 is supported by four springs 81 to float above the silicon substrate. One end of each spring 81 is fixed to a corresponding one of a pair of spring bases 80 fixed to the silicon substrate, and the entire spring 81 is separated from the silicon substrate.
The stationary elements 83 and 84 are respectively made up of main body portions 83a and 84a, and support portions 83b and 84b constituting leg portions vertically extending from the centers of the main body portions 83a and 84a. The end portions of the support portions 83b and 84b oppose each other. Many comb tooth portions 91 are formed in a comb tooth shape at a predetermined pitch in two lines on the two sides of each of the support portions 83b and 84b. As shown in FIG. 13, many stationary element electrodes 93 are formed at a predetermined pitch in a comb tooth shape on one side of each comb tooth portion 91.
The movable element 82 is made up of a pair of parallel support portions 82a and a coupling portion 82b coupling the centers of the support portions 82a. The movable element 82 is combined with the stationary elements 83 and 84 to constitute an actuator. That is, the support portions 82a of the movable element 82 are arranged parallel to sandwich the support portions 83b and 84b of the stationary elements 83 and 84. The coupling portion 82b of the movable element 82 vertically crosses the gap formed by the end portions of the support portions 83b and 84b of the stationary elements 83 and 84.
The movable element 82 comprises many comb tooth portions 92 formed in a comb tooth shape at the same pitch as that between the comb tooth portions 91 of the stationary elements 83 and 84. The comb tooth portions 91 of the stationary elements 83 and 84 and the comb tooth portions 92 of the movable element 82 overlap and interdigitated with each other. As shown in FIG. 13, movable element electrodes 94 to be inserted between the stationary element electrodes 93 are formed on one side of each comb tooth portion 92. The comb tooth portion 91 of each of the stationary elements 83 and 84 has a larger width than that of the comb tooth portion 92 of the movable element 82. The stationary element electrode 93 has a larger width than that of the movable element electrode 94.
As shown in FIG. 14, the comb tooth portion 91 formed integrally with the stationary element electrode 93 is fixed to a silicon substrate 100 via a stationary element base 101. In contrast to this, the comb tooth portion 92 formed integrally with the movable element electrode 94 is separated from the silicon substrate 100, i.e., floats above the surface of the semiconductor substrate 100 at a predetermined interval.
In this arrangement, the movable element 82 can be moved right or left in FIG. 12, i.e., the comb tooth portion 92 can be moved in a direction to come close to and separate from the comb tooth portions 91 by applying a voltage across the movable element electrode 94 of the comb tooth portion 92 and the stationary element electrodes 93 of the stationary elements 83 and 84. In this case, the movable element 82 can be moved left by applying a voltage to the left stationary element 84, or right by applying a voltage to the right stationary element 83.
A method of manufacturing the microactuator having this arrangement will be explained. A 2-.mu.m thick PSG (PhoshoSilicate Glass) film is patterned in a region on the silicon substrate 100 where the movable element 82 is to be formed. Copper is plated between resist patterns formed on the PSG film using photolithography.
The PSG film is removed using hydrofluoric acid to separate the movable element 82 including the movable element electrode 94 from the silicon substrate 100, thereby forming the copper-plated movable element 82. In this way, the microactuator in reference 1 using a 20-.mu.m thick copper material is manufactured.
In a microactuator using a silicon IC process, a structure using a polysilicon thin film has conventionally been known well. Compared to the electroplated actuator, the microactuator with a polysilicon structure has good matching with the silicon IC process and exhibits excellent mechanical characteristics. Note that in applications to a magnetic/magneto-optical head and the like, movement of the head in directions other than a desired direction must be suppressed small.
In the microactuator shown in FIG. 12, the movable element 82 must move right and left in FIG. 12, but its movement in a direction perpendicular to the surface of the silicon substrate 100 must be suppressed as small as possible. From this condition, the spring 81 must be made thick. The movable element electrode 94 and the stationary element electrode 93 must also be made thick in order to use a large electrostatic force.
From these conditions, a microactuator having an electrode thickness of 20 .mu.m or more must be manufactured for practical use. Since the polysilicon thin film has a thickness of about 4 .mu.m at most, microactuators using the above-described plating technique and a single-crystal silicon etching technique (to be described later) are being developed.
To manufacture a microactuator made of single-crystal silicon, the method using an SOI (Silicon On Insulator) substrate described in A. Benitez et al., "Bulk Silicon Microelectromechanical Devices Fabricated from Commercial Bonded and Etched-Back Silicon-on-Insulator Substrates", Sensors and Actuators, A50, pp. 99-103, 1995 (reference 2) can be employed.
According to this method, the movable element electrode 94 and the stationary element electrode 93 in FIG. 14 are formed of a 20-.mu.m thick single-crystal silicon film, and the stationary element base 101 is formed of a silicon oxide film. By removing the silicon oxide film positioned below the movable element electrode 94 using hydrofluoric acid, the movable element electrode 94 can be separated from the silicon substrate 100.
In this case, since the movable element electrode 94 is narrower in width than the stationary element electrode 93, the silicon oxide film is still left below the stationary element electrode 93 even upon etching using hydrofluoric acid, and forms the stationary element base 101. In this manner, the movable element electrode 94 and the stationary element electrode 93 each made of, e.g., a 20-.mu.m thick single-crystal silicon film are formed on the silicon substrate 100.
In the microactuator with a predetermined electrode thickness manufactured in the above manner, if the movable element electrode 94 collides against and attaches to the silicon substrate 100 due to some reason during driving, the movable element 82 stops operating. This typically occurs when the movable element electrode 94 and the stationary element electrode 93 are formed using single-crystal silicon because both the movable element electrode 94 and the silicon substrate 100 have mirror surfaces. The above problem is also frequently posed when the movable element electrode 94 and the stationary element electrode 93 are formed using copper plating.
Recently, the method of effectively solving the above problem is reported in association with a polysilicon microactuator in Y. Yee et al., "Polysilicon Surface Modification Technique to Reduce Sticking of Microstructures", Digest of The 8th International Conference on Solid-State Sensors and Actuators, Vol. 1, pp. 206-209, June 1995 (reference 3).
A method of manufacturing an actuator using this method of solution will be described with reference to FIGS. 15A to 15E.
After an oxide film 121 is formed on a silicon substrate 120, a 0.5-.mu.m thick polysilicon film 122 is formed on the oxide film 121 (FIG. 15A). A PSG oxide film 123 is formed on the polysilicon film 122 (FIG. 15B). At this time, oxidization progresses deep in the grain region of the polysilicon film 122. This sample is dry-etched without any mask to increase differences in thickness of the PSG oxide film 123 and form a polysilicon film 124 having microstructures on the surface (FIG. 15C).
A 2-.mu.m thick PSG film 126 is formed on the polysilicon film 124, and a 2-em thick polysilicon film 125 is formed on the PSG film 126. The polysilicon film 125 is patterned to have a microactuator electrode shape (FIG. 15D). Finally, the PSG film 126 is removed using hydrofluoric acid to separate an electrode made of the polysilicon film 125 from the surface of the silicon substrate 120 (FIG. 15E).
According to the method described above, the surface of the silicon substrate 120 in contact with the movable element electrode is roughened to decrease the two contact areas, thereby decreasing the attraction force between the solid-state surfaces. The method of reducing friction by forming microstructures on the two contact surfaces has conventionally been reported. For example, the method of arranging projections of an oxide film on the surface of a substrate opposing a movable element is disclosed in Japanese Patent Laid-Open No. 8-23685, "Electrostatic Microactuator" (reference 4).
However, since the projections of the oxide film described in reference 4 are formed by photolithography, the planar size of the projection is much larger than that in the method of roughening the surface, described with reference to FIGS. 15A to 15E. The friction or the attaching force sensitively depends on the degree of surface roughness, so that the characteristics of the microactuator by the manufacturing method described with reference to FIGS. 15A to 15E are superior to the characteristics of the actuator having the conventional projection structure. Further, since the method in FIGS. 15A to 15E does not require photolithography to form microstructures, the whole process can be simplified.
However, the conventional method shown in FIGS. 15A to 15E cannot be applied to the manufacture of a single-crystal silicon actuator using an SOI wafer owing to the following reason.
More specifically, in the manufacturing method using the SOI wafer, an SOI wafer purchased from a semiconductor wafer manufacturer is normally used. This is because the cost is low and the substrate quality is excellent. When a microactuator is to be manufactured using the bought SOI wafer, however, it is essentially impossible to roughen the substrate surface below the movable element using the manufacturing method shown in FIGS. 15A to 15E because films used for the movable element and the stationary element have already been formed.
The conventional microactuator also has the following structural problem. During operation, the movable element attaches to not only the substrate but also the stationary element because the movable element electrodes 94 and the stationary element electrodes 93 complicatedly interdigitated with each other, as shown in FIG. 13. As the electrodes 93 and 94 become thicker, and the interval between the electrodes 93 and 94 becomes narrower, the above problem more frequently occurs.
This problem of attaching the movable element electrode 94 to the stationary element electrode 93 poses common problems in not only the microactuator using the SOI wafer but also all microactuators manufactured by polysilicon and electroplating.
This problem cannot be satisfactorily solved by the manufacturing method described with reference to FIGS. 15A to 15E. This is because the microstructure 124 is formed by dry etching using a plasma aligned in a direction perpendicular to the surface of the silicon substrate 120 in etching the polysilicon film 122 by the method in FIGS. 15A to 15E. That is, this method can be applied to only processing of a surface perpendicular to the silicon substrate 120.
When the movable element electrode is to be formed of the polysilicon film 122, compression stress acts on the polysilicon film 122 by the oxidization step in FIG. 15B. For this reason, the electrode itself separated from the silicon substrate 120 deforms.